Feature generation based on relation learning and image partition for occluded person re-identification

In order to solve the challenging tasks of person re-identification(Re-ID) in occluded scenarios, we propose a novel approach which divides local units by forming high-level semantic information of pedestrians and generates features of occluded parts. The approach uses CNN and pose estimation to extract the feature map and key points, and a graph convolutional network to learn the relation of key points. Specifically, we design a Generating Local Part (GLP) module to divide the feature map into different units. Based on different occluded conditions, the partition mode of GLP has high flexibility and variability. The features of the non-occluded parts are clustered into an intermediate node, and then the spatially correlated features of the occluded parts are generated according to the de-clustering operation. We conduct experiments on both the occluded and the holistic datasets to demonstrate its effectiveness.     

A matching game framework for users clustering and resource allocation with wireless power transfer in a CR-NOMA network

This paper investigates the resource allocation in a massively deployed user cognitive radio enabled non-orthogonal multiple access (CR-NOMA) network considering the downlink scenario. The system performance deteriorates with the number of users who are experiencing similar channel characteristics from the base station (BS) in NOMA. To address this challenge, we propose a framework for maximizing the system throughput that is based on one-to-one matching game theory integrated with the machine learning technique. The proposed approach is decomposed to solve users clustering and power allocation subproblems. The selection of optimal cluster heads (CHs) and their associated cluster members is based on Gale-Shapley matching game theoretical model with the application of Hungarian method. The CHs can harvest energy from the BS and transfer their surplus power to the primary user (PU) through wireless power transfer. In return, they are allowed to access the licensed band for secondary transmission. The power allocation to the users intended for power conservation at CHs is formulated as a probabilistic constraint, which is then solved by employing the support vector machine (SVM) algorithm. The simulation results demonstrate the efficacy of our proposed schemes that enable the CHs to transfer the residual power while ensuring maximum system throughput. The effects of different parameters on the performance are also studied.     

一种实时多尺度目标检测识别算法

无人机目标检测与识别任务中,目标随着飞行高度的改变尺寸发生显著变化。常规目标检测模型中,获取的小目标细节信息有限,检测精度较低;而适用于小目标的实时检测模型往往容易丢失大目标的背景信息,降低大目标的检测精度。针对以上多尺度目标检测识别任务难点,提出一种基于改进特征金字塔网络(Feature Pyramid Network, FPN)结构的实时多尺度目标检测识别模型。该模型通过增加特征金字塔层级覆盖更广的目标尺度,获取更为丰富的目标信息;同时,利用跨连接增加不同尺度特征融合的多样性,降低特征传导距离,保留更加完整的尺度特征来提高模型检测识别多尺度目标的性能。通过实验发现,相比于原始网络结构和相同特征层级的四层特征金字塔结构,加入改进特征金字塔结构的多尺度目标检测模型识别性能得到了提升。     

Approximate minimum-cost multicommodity flows in\tilde O(ɛ −2 KNM) timetime

We show that an ε-approximate solution of the cost-constrainedK-commodity flow problem on anN-nodeM-arc network,G can be computed by sequentially solving O(K(? ^?2+logGK) logGM log (G? ^?1 GK)) single-commodity minimum-cost flow problems on the same network. In particular, an approximate minimum-cost multicommodity flow can be computed in $\tilde O$ (G? ^?2 GKNM) running time, where the notation Õ(·) means “up to logarithmic factors”. This result improves the time bound mentioned by Grigoriadis and Khachiyan [4] by a factor ofM/N and that developed more recently by Karger and Plotkin [8] by a factor of? ^?1. We also provide a simple $\tilde O$ (NM)-time algorithm for single-commodity budget-constrained minimum-cost flows which is $\tilde O$ (? ^?3) times faster than the algorithm developed in the latter paper.     

Immobilization of α-amylase in vinyl-polymer-based interpenetrating polymer networks

Semi-interpenetrating polymer networks (IPNs) of poly(ethylene glycol), poly(vinyl alcohol) and polyacrylamide were prepared as a support for enzyme immobilization and kinetic studies were performed for the immobilization of -amylase. The effect of IPN composition on the extent of immobilization was investigated and the percentage of relative activity of the immobilized enzyme was evaluated as a function of the chemical architecture of the IPNs, pH and temperature, taking starch as a substrate. The kinetic constants and the maximum reaction velocity were also evaluated. The IPNs were characterized by IR spectral analysis.     

DWDM上25.344 Mbit/s光监控信息传送系统的设计

采用25.344 Mbit/s光监控信息传送技术可以更好地满足当前密集波分复用(DWDM)系统中不断增加的网络监控信息和附加业务的需要.文章首先介绍了25.344 Mbit/s光监控信息传送系统的技术方案,然后详细描述了其帧结构、码型设计以及硬件实现的框图,最后介绍了该技术的优点.     

基于智能网的变声彩话业务的设计与实现

彩话业务(CTS)是继彩铃业务之后电信运营商推出的又一语音增值业务,包括普通彩话业务(oCTS)和变声彩话业务(dCTS)两种.本文首先简要介绍了dCTS的概念及业务属性,然后从系统架构和业务流程两个方面介绍了基于智能网的dCTS的设计和实现方案,最后对本业务的商用前景及局限性进行了讨论.     

特斯拉线圈电网络模型的宽频等值化简

针对特斯拉线圈复杂的电网络模型,提出了利用求解非线性方程和优化目标函数建立等值网络的方法。该等值网络在宽频范围内,在阻抗参数、有载增益、输入阻抗、传输功率与传输效率方面,均与原网络近似等值。该方法应用电压、电流之积计算灵敏度,方便地获得雅可比矩阵和目标函数梯度。本文方法对“网络分析与综合”课程或“电路理论”课程的师生,以及科研人员,具有切实的启发作用。     

水体透射光谱结合主成分分析(PCA )改进化学需氧量(COD)含量估算研究

化学需氧量（Chemical Oxygen Demand,COD）是水体有机污染的一项重要指标,化学需氧量越高,表示水污染程度越严重。 为了解决传统的COD测量方法耗时较长,不利于快速、实时地获取水体中COD的信息等问题。本文提出了基于透射光谱测量结合主成分分析(Principal Component Analysis, PCA)改进水体COD含量估算模型。具体的,采集100组COD水体光谱信息,分别使用3种不同的高光谱数据预处理方法对光谱数据进行预处理,分析不同预处理方法对模型精度的影响,并基于不同的预处理方法分别建立高斯过程回归模型(Gaussian Process Regression, GPR)和BP神经网络模型,分析不同预处理方法对模型精度的影响;并对各模型结合PCA数据降维方法进行模型的改进,通过比较模型的精度选择最优模型进行水体COD含量的检测。结果显示,相比于原始光谱数据建立的GPR模型和BP神经网络模型,数据预处理后的模型精度明显提升;且结合PCA对预处理后的数据进一步降维处理后,模型精度得到了进一步的提升。其中,基于标准正态变量变换特征结合PCA改进BP神经网络模型基于PCA改进的BP神经网络模型R^2高达0.9940,均方根误差RMSE为0.022540。证明了基于PCA改进的BP神经网络数据降维方法对预处理后的光谱数据进行降维处理,有利于去除光谱中的冗余信息,提取特征信息,可以实现高光谱检测方法可以实现COD含量估算模型的优化,从而为传统COD测量方法存在的问题提出了一种新的解决思路。     

An Extensive Assessment of Network Embedding in PPI Network Alignment

Network alignment is a fundamental task in network analysis. In the biological field, where the protein–protein interaction (PPI) is represented as a graph, network alignment allowed the discovery of underlying biological knowledge such as conserved evolutionary pathways and functionally conserved proteins throughout different species. A recent trend in network science concerns network embedding, i.e., the modelling of nodes in a network as a low-dimensional feature vector. In this survey, we present an overview of current PPI network embedding alignment methods, a comparison among them, and a comparison to classical PPI network alignment algorithms. The results of this comparison highlight that: (i) only five network embeddings for network alignment algorithms have been applied in the biological context, whereas the literature presents several classical network alignment algorithms; (ii) there is a need for developing an evaluation framework that may enable a unified comparison between different algorithms; (iii) the majority of the proposed algorithms perform network embedding through matrix factorization-based techniques; (iv) three out of five algorithms leverage external biological resources, while the remaining two are designed for domain agnostic network alignment and tested on PPI networks; (v) two algorithms out of three are stated to perform multi-network alignment, while the remaining perform pairwise network alignment.